Fitness effects of inbreeding and outbreeding on
golden paintbrush (Castilleja levisecta):
Implications for recovery and reintroduction
September 2003
Thomas N. Kaye
Beth Lawrence
Institute for Applied Ecology
Corvallis, Oregon
Washington Department of Natural Resources
and
Institute for Applied Ecology
PREFACE
This report is the result research conducted by the Institute for Applied Ecology (IAE) under contract
with a public agency, the Washington Department of Natural Resources, Natural heritage Program
under Contract No. 03-218. IAE is a non-profit organization dedicated to natural resource conservation,
research, and education. Our aim is to provide a service to public and private agencies and individuals
by developing and communicating information on ecosystems, species, and effective management
strategies and by conducting research, monitoring, and experiments. IAE offers educational
opportunities through 3-4 month internships. Our current activities are concentrated on rare and
endangered plants, habitat management and invasive species.
Questions regarding this report or IAE should be directed to:
Thomas N. Kaye
Institute for Applied Ecology
227 SW 6th
Corvallis, Oregon 97333
phone: 541-753-3099
fax: 541-753-3098
ACKNOWLEDGMENTS
Funding for this project was provided by the Washington Department of Natural Resources
and Institute for Applied Ecology. We thank the Oregon State University Seed Lab and Dale
Brown for providing germination assistance and facilities used in this study. We also thank
Melissa Carr, Emerin Hatfield, Rebecca Kessler, Melanie Barnes, and Angela Brandt
(IAE/Native Plant Society of Oregon Interns) for assistance with plant propagation and data
collection. Manuela Huso provided generous and invaluable statistical support and assistance.
We thank Florence Caplow and Carla Wise for their discussions and comments on an earlier
draft of this report.
Inbreeding and outbreeding in Castilleja levisecta, 2003
i
TABLE OF CONTENTS
Preface
i
Acknowledgments
i
Introduction
Background
The importance of mating system: inbreeding and outbreeding
1
1
1
Methods
Pollination experiment
Propagation of crossed progeny
Analysis of fitness measures
4
4
6
6
Results
Effects of inbreeding and outbreeding on fitness
8
8
Discussion
Mating system of Castilleja levisecta
Effects of inbreeding and outbreeding on fitness
Recommendations for recovery and reintroduction
11
11
11
14
Literature cited
16
Inbreeding and outbreeding in Castilleja levisecta, 2003
ii
INTRODUCTION
Background
Castilleja levisecta (golden paintbrush)
once occupied prairies and grasslands
throughout the Puget Trough and
Willamette Valley. Habitat destruction
and alteration over the past century have
resulted in substantial declines in native
vegetation in this ecoregion, and several
species, including golden paintbrush, are
now listed by state and federal agencies as
threatened or endangered. All remaining Figure 1. Castilleja levisecta (golden paintbrush).
populations of golden paintbrush occur in
Washington and British Columbia; the species is considered to be extirpated in Oregon.
Castilleja levisecta is an herbaceous perennial that appears to reproduce only by seed. Many
populations are declining, and past research indicates that disturbances such as fire and
mowing may be useful for maintaining or expanding existing populations (USFWS 2000).
Like most paintbrushes (Heckard 1962), this species is a hemiparasite – its roots penetrate the
roots of neighboring plant species and derive nutrients, carbohydrates, and possibly other
secondary compounds from these hosts.
The Recovery Plan for golden paintbrush (USFWS 2000) identifies population reintroduction,
development of propagation methods, and studies of the pollination biology of the species as
high priority actions to meet recovery objectives. Because no populations of this species are
known to remain in Oregon, population reintroduction will be crucial to recovery in this region
as well as portions of the range to the north.
Inbreeding and outbreeding in Castilleja levisecta, 2003
1
The importance of mating system: inbreeding and outbreeding
Creating new populations from propagated material should consider the source of these
materials as well as their genetic relationship to one another. When plants self or mate with
close relatives their offspring may show negative effects related to inbreeding depression. On
the other hand, individuals that are distantly related may produce offspring that show higher
fitness than either parent, or their offspring can have reduced vigor due to outbreeding
depression (see reviews by Kaye 2001a and Hufford and Mazer 2003). The reintroduction
plan for C. levisecta (Caplow 2002) recommends that the risks of inbreeding and outbreeding
depression be examined experimentally.
Inbreeding depression – Inbreeding depression occurs when close relatives mate (or plants
self-fertilize) and their offspring display reduced vigor or fitness. Inbreeding depression is a
well-known and studied phenomenon, and often occurs in small, fragmented, or isolated
populations, or when mating is frequent between close neighbors. It may result when
deleterious recessive alleles are paired (creating homozygotes) so that their negative effects are
expressed in the progeny. When these genes are not paired (as after outcrossing), they may be
masked by a more favorable allele (as a heterozygote), so the progeny function normally.
Inbreeding depression may also result from loss of heterozygote advantage. In plants,
inbreeding depression can be expressed at any stage in the life cycle, including seed
germination, seedling establishment, plant growth rate and survival, flowering, and seed
production. Populations suffering from inbreeding depression can often benefit from
out-crossing with individuals in other populations, which may result in higher heterozygosity,
improved health of individuals, and greater population viability. This is one factor used to
support the use of multiple sources of plant materials in restoration (one side of the Single or
Multiple Source debate [Kaye 2001a]).
Outbreeding depression – Outbreeding depression, which is a reduction in fitness of progeny
from distant parents, has a much shorter history of study and is less documented and
understood than inbreeding depression. A recent literature search (Kaye 2001a) found 468
papers on inbreeding depression but only 25 references to outbreeding depression. Even so,
this hot topic in genetic and conservation research has been demonstrated in various organisms,
including salmon (Gharrett 1999), fruit flies (Aspi 2000), and chimpanzees (Morin et al.
Inbreeding and outbreeding in Castilleja levisecta, 2003
2
1992). Some animal studies have found a positive effect of outbreeding, however, such as in
bats (Rossiter et al. 2001). Among plants it may occur in larkspur (Waser and Price 1991,
1994), skyrocket (Waser et al. 2000), a carnivorous pitcher plant (Sheridan and Karowe
2000), Hawaiian silversword (Friar et al. 2001), a Mediterranean borage (Quilichini et al.
2001), a subshrub (Montalvo and Ellstrand 2001), and an exotic roadside weed (Keller et al.
2000). In many cases, crossing between unrelated individuals results in F1 progeny with
increased fitness, followed by the expression of outbreeding depression in later generations
(Hufford and Mazer 2003). Most researchers (e.g., Lynch 1991, Waser 1993) believe that
there is hybrid vigor in the first generation followed by reduced fitness in later generations
from loss of ecological adaptation (at least one of the original parents was poorly adapted to
the site) and/or disruption of coadapted gene complexes.
In this paper we examine the mating system of Castilleja levisecta and explore the effects of
inbreeding and outbreeding on F1 progeny fitness at multiple life history stages. The results
have direct bearing on reintroduction and recovery actions for the species because the risks
associated with inbreeding and outbreeding need to be identified in order to assist with seed
selection and genetic management of outplanting activities.
Inbreeding and outbreeding in Castilleja levisecta, 2003
3
METHODS
Pollination experiment
We performed controlled cross-pollination between
individuals of differing known heritage. These crosses were
of four major types: self-pollination, crosses between
siblings, crosses between non-sibling plants from the same
source population, and crosses between individuals from
different populations. The plants used in these crosses were
grown in pots from seed collected in wild populations on
Whidbey Island and San Juan Island, Washington in
September, 2000. Sources of plants and the number of each
type of cross are listed in Table 1. Genetic identities among
Figure 2. The inflorescence of
Castilleja levisecta is a raceme, and
the three source sites (False Bay [San Juan Island], West
the individual flowers arepartly
Beach, and Fort Casey [Whidbey Island]) were all very high covered by a showy yellow bract
(shown here pulled back to expose a
(0.93-0.94, Godt and Hamrick [2002]) although their
flower). The receptive stigmas can be
geographic distances ranged from 13.3 to 45.4 km (Table 1). seen as two-lobed, rounded structures
emerging above the flowers.
Seeds from individual plants were kept separate at the time of
collection, so that maternal lines could be documented.
Crosses between siblings involved pollination between two
plants grown from seed taken from the same maternal parent.
Flowers of Castilleja levisecta are borne on a raceme, and
each flower has one pistil and stigma as well as a single
ovary with numerous ovules. The mean number of ovules
per flower in C. levisecta is 183 ± 5.8 (Kaye, unpubl. data).
Each fruit is a capsule and the seeds are typically very small
(~0.5 mm). The flowers are protogynous; their pistils
extend beyond the opening of the flower and the stigma
becomes receptive prior to anther dehiscence (Figure 2). We
grew all plants in a screen house to prevent insects from
visiting the experimental flowers and moving pollen between
plants.
Inbreeding and outbreeding in Castilleja levisecta, 2003
Figure 3. Pollinated flowers of C.
levisecta were marked with colorcoded thread.
4
To prevent unwanted self-pollination, all crosses first involved emasculation, i.e., removing
the anthers from the flowers identified as pollen recipients (dames) prior to anther dehiscence.
We then used forceps to remove a mature, dehiscing anther from a pollen donor plant (sire)
and applied pollen directly to the stigma of the dame. Each dame flower on a raceme was
marked with color coded thread tied around its base (Figure 3).
When fruits matured, they were removed from the dames and returned to the lab for
examination under a dissecting microscope. Each fruit was measured for length and width,
then opened to count normally developed seeds as well as undeveloped ovules. We used this
information to calculate the proportion of total ovules that set seed.
Table 1. Crossing types, source populations involved, and sample sizes (number of crosses
conducted). Geographic distances between populations are given in parentheses.
Cross type
source populations used
sample size
self
Fort Casey
2
West Beach
17
Total=19
sibling
False Bay
3
West Beach
9
Total=12
within-population
False Bay
5
Fort Casey
7
West Beach
23
Total=35
between-population
False Bay × Fort Casey (45.4 km)
6
False Bay × West Beach (33.1 km)
9
Fort Casey × West Beach (13.3 km)
12
Total=27
Inbreeding and outbreeding in Castilleja levisecta, 2003
5
Propagation of crossed progeny
To germinate seeds from each cross, the seeds were placed on germination pads moistened
with distilled water in individual plastic boxes with tight fitting lids. We used 50 seeds from
each cross except where fewer were available. The seeds were then exposed to cool
temperatures (4 °C) for six weeks, followed by alternating warm temperatures (15 °C and 25
°C) for two weeks. This method has resulted in high germination rates (generally >75%),
depending on source population) in previous studies with C. levisecta (Wentworth 1998, Kaye
2001b). After these treatments we counted the total number of seedlings emerging in each
box.
Up to ten seedlings from each cross were then planted in fine potting soil in 5 x 5 cm by 6 cm
deep pots in a greenhouse. The plants were watered from beneath by flooding the greenhouse
bench for 30 minutes, twice per week. After 4 weeks of growth, the small plants were then
repotted into 10 x 10 cm by 9 cm deep pots with a seedling of Eriophyllum lanatum, an
herbaceous plant in the Asteraceae that proved to be a superior host in earlier propagation
experiments with C. levisecta (Kaye 2001b). The potted plants were then moved out of doors
to a flat garden bed and allowed to grow an additional 8 weeks. All plants were fertilized once
per week throughout this period. The total growth period from first potting to the end of the
experiment was 89 days (ending 30 June 2003). We measured plant size on the 89th day as
number of stems, length of each stem, and number of flowering racemes.
Analysis of fitness measures (seed set, seed germination, plant growth, flowering)
To determine the effect of crossing distance on fitness in Castilleja levisecta, we performed
separate statistical analyses on each measure of plant performance we observed (Table 2). We
tested the hypothesis that pollination-type had no effect on seed set with a Kruskal-Wallis
one-way ANOVA on ranks combined with Kruskal-Wallis multiple-comparison Z-value tests
to compare individual treatments. These non-parametric tests were used because the seed-set
data, especially residuals for sibling crosses, did not conform to the assumptions of normal
distribution and equal variances required for standard ANOVA. We examined seed
germination and flowering with logistic regression to determine if the odds of germination or
Inbreeding and outbreeding in Castilleja levisecta, 2003
6
flowering differed among the treatments. Our examinations of these data suggested that their
distributions were not strictly binary, so we used quasi-likelihood methods. And finally, we
used one-way ANOVA to test for a treatment effect on plant growth expressed as the sum of
all stem lengths on each plant, followed by Fisher’s protected LSD multiple-comparison test to
compare individual treatments.
Table 2. Fitness measures used to evaluate the effects of inbreeding and outbreeding on
Castilleja levisecta, the statistical analysis implemented in this study, and sample sizes for each
treatment.
Fitness measure (response variable)
Statistical analysis
Seed set (proportion of ovules developing
Kruskal-Wallis one-way ANOVA on ranks
into normal seeds)
with Kruskal-Wallis multiple-comparison
Z-value test. Sample sizes: self=19,
sibling=12, same population=35, different
population=27.
Seed germination (odds of seeds
Logistic regression, quasi-likelihood.
germinating)
Sample sizes: self=8, sibling=10, same
population=35, different population=26.
Plant growth (total stem length)
One-way ANOVA, Fisher’s protected LSD
multiple-comparison test. Sample sizes:
self=6, sibling=9, same population=35,
different population=25.
Flowering (odds of producing a flowering
Logistic regression, quasi-likelihood.
stem)
Sample sizes: self=6, sibling=9, same
population=35, different population=25.
Inbreeding and outbreeding in Castilleja levisecta, 2003
7
RESULTS
Effects of inbreeding and outbreeding on fitness
Seed set – Castilleja levisecta appears to be almost completely self-incompatible, and seed-set
appears to increase as the genetic relationship among mating individuals becomes more distant.
The proportion of ovules that set normal seeds was significantly affected by pollination crosstype in Kruskal-Wallis ANOVA (df=3, P2 corrected for ties=44.3, P<0.0001). Selfpollinations had the lowest seed-set, averaging 0.7%, which was significantly less than the
average 33% ± 7% (SE) for sibling crosses (Figure 4). Crosses between individuals from
different populations had much higher seed-set (0=80% ± 5%) than those between siblings,
while matings between non-sibling plants from the same population had intermediate seed set
(0=71% ± 4%).
90%
c
bc
80%
70%
viable seed set
60%
50%
b
40%
30%
20%
10%
a
0%
self
sibling
same population
different population
pollen source
Figure 4. Mean (±1 SE) percentage seed-set in Castilleja levisecta from different pollination cross-types. Bars
with the same letter do not differ at the "=0.05 level of probability.
Inbreeding and outbreeding in Castilleja levisecta, 2003
8
Seed germination – We found no evidence that seed germination was affected by cross type.
Logistic regression did not detect any significant difference in germination rate among the
mating types tested here (df=3,75; P2=0.32; P=0.957). The overall seed germination rate
averaged 84.5% ± 3% (SE). Note that few seeds were produced from self and sibling
crosses, so that little statistical power was available to detect differences between these matings
and other types. This rate of germination is typical of wild-collected seeds, which tend to have
greater than 75% germination (Kaye 2001b) under the same environmental conditions as
applied in the current study.
Plant Growth – Plant size was affected substantially by mating type (df=3,71; F=0.12.72;
P<0.00001). F1 plants from self-pollinations grew an average of 55 cm ± 19 cm (n=6) of
stem in 89 days, and were not significantly different from plants resulting from crosses
between siblings (0=79 cm ± 16 cm SE, n=9) (Figure 5). However, both of these cross
types produced plants significantly smaller than non-sibling crosses in the same population
(0=130 cm ± 8 cm SE, n=35), which were in turn significantly smaller than plants from
between-population crosses (0=162 cm ± 9 cm SE, n=25).
200
c
180
stem production (cm)
160
b
140
120
a
100
80
a
60
40
20
0
self
sibling
same population
different population
pollen source
Figure 5. Mean (±1 SE) plant size (total stem production) of Castilleja levisecta plants grown from seed
produced by different pollination types. Bars with the same letter do not differ at the "=0.05 level of
probability.
Inbreeding and outbreeding in Castilleja levisecta, 2003
9
Flowering – The rate at which plants flowered was also significantly different among plants
from different cross-types (logistic regression; df=3,71; P2=8.77; P=0.033). On average,
11% (±3% SE) of crosses between non-sibling individuals from the same population produced
offspring that flowered (Figure 6), and the odds of these plants flowering was not significantly
different from self or sibling crosses. However, the odds of flowering for plants from crosses
between different populations was 1.1 to 4.6 times higher than within population crosses.
30%
25%
flowering rate
20%
15%
10%
5%
0%
self
sibling
same population
different population
pollen source
Figure 6. Mean (±1 SE) percentage flowering of F1 plants from different cross types.
Inbreeding and outbreeding in Castilleja levisecta, 2003
10
DISCUSSION
Mating system of Castilleja levisecta
Castilleja levisecta appears to require out-crossing for reliable seed set because the species
possesses barriers to self-fertilization. These barriers include protogyny in the flowers which
may prohibit self-pollination in the field (but possibly not geitonogamy by insects), as well as
genetic or physiological self-incompatibility. Although the mechanism of self-incompatibility
is unknown, it appears to block seed set almost completely, with the exception that a small
number of seeds may be produced in fruits from self-pollinations. We found that among 19
selfings, only 8 produced fruits with any filled seeds and these seeds generally numbered only
1 or 2, while in outcrossed plants fruits containing seeds were produced from all pollinations.
Self-incompatibility may be typical in the genus Castilleja. Although we found no published
studies that specifically address compatibility within species of this genus, Heckard (1968), in
his extensive study of polyploidy in Castilleja, stated that most species appear to be self-sterile
and no seed is produced unless pollinators are present.
Many species of plants maintain a mixed-mating system in which both self- and crosspollination serve to fertilize embryos. However, because Castilleja levisecta is nearly selfincompatible, the vast majority of seeds produced in its populations are likely from insectmediated crosses between different individuals. This prediction is consistent with findings
from a recent genetic study (Godt and Hamrick 2001) of allozyme diversity in C. levisecta,
which found no evidence for inbreeding in its populations. The authors also noted that there
could be strong selection against inbred individuals and our measurements of fitness traits of
inbred plants support this assumption. In addition, heterozygosity is generally higher than
expected in populations of C. levisecta (Godt and Hamrick 2001), and self-incompatibility and
failure of flowering in inbred individuals could have contributed to this.
Effects of inbreeding and outbreeding on fitness
We found evidence for effects of crossing distance on fitness of Castilleja levisecta at multiple
stages in the life cycle of the species. Both inbreeding depression and outbreeding advantage
Inbreeding and outbreeding in Castilleja levisecta, 2003
11
(hybrid vigor) were detected. For example,
inbreeding, including selfing and inbreeding
among siblings (which, on average, share 50%
of their alleles), reduced seed production and
plant growth of the F1 generation significantly
when compared to crosses between non-siblings
from the same or different populations (Figures
4, 5 and 7). Inbreeding depression was not
detected in seed germination or flowering rate,
but this may be because of insufficient statistical Figure 7. Castilleja levisecta individuals (30 days
old) from self-pollination (left) and outbreeding
power due to the small sample sizes available
for the inbred treatments.
(right).
One recent example of inbreeding depression (Richards 2000) in a perennial plant, white
campion (Silene alba), showed that isolated populations had high inbreeding depression (in the
form of low seed germination success), crosses between related individuals resulted in reduced
germination success, and gene-flow was higher between unrelated individuals. This study is
important because it demonstrates the potential for a “rescue-effect” for populations
experiencing inbreeding depression by intentionally mixing unrelated individuals into such a
population. Although inbreeding has not been detected in Castilleja levisecta populations to
date (Godt and Hamrick 2002), small populations could experience a decline in seed
production due to insufficient numbers of genetically compatible individuals. In addition,
inbreeding depression could accumulate in very small populations in which the surviving
members are genetically related. In either case, intentionally introducing pollen or individual
plants from other populations (preferably nearby sites) could improve seed production and play
a role in population recovery.
Outbreeding depression was not detected at any level in our study of seed set and F1 plant
traits. Instead, we found significant increases in fitness of plants produced by crosses of
individuals from different populations. In particular, plant growth and flowering rate were
substantially higher in outbred plants (Figures 5 and 6). Crosses between plants from different
populations produced F1 individuals that produced 25% more stem length, on average, than
Inbreeding and outbreeding in Castilleja levisecta, 2003
12
plants from within-population crosses. Further, flowering was 1.1 to 4.6 times more likely in
outbred plants than those from within population matings.
One concern about this study is that multiple populations were used for each treatment and
pooled for analysis. For example, sibling crosses involved two different source populations,
and their relative decline in fitness might best be compared only to their parents or other plants
from the same population. One reason for this concern is that differences in heterozygosity
among populations could affect the severity of inbreeding depression detected in sibling
crosses. However, we believe that our analysis with pooled data is justified because the levels
of heterozygosity in these populations were similar (and high, 0.174 at West Beach and 0.231
at False Bay; Godt and Hamrick 2002). In addition, measurements of plant fitness of parental
plants of all crosses detected no significant differences in stem production or flowering (Kaye,
unpubl. data). Another concern is that the crosses between populations to detect outbreeding
effects could differ if the various pairs of populations crossed differed in their genetic
similarity. For example, a between-population cross between two populations that were
similar genetically might result in a different effect than a cross between two genetically
distant populations. The pairs of populations used for crosses in our study all had high and
similar genetic identities (0.93 - 0.94, Hamrick and Godt 2002), despite variability in their
geographic distances (Table 2).
Although we did not detect outbreeding depression in the F1 generation, this phenomenon may
yet act in Castilleja levisecta. Several studies (reviewed in Kaye 2001a and Hufford and
Mazer 2003) have found that expression of outbreeding depression may be delayed to the F2
or F3 generations. Even so, it may not occur at all, and some studies have found no risk of
outbreeding depression even for very long-distance crosses. For example, in a study of
partridge pea (Chamaecrista fasciculata, an annual legume) Fenster and Galloway (2000)
collected plants from various populations ranging from 100 m to 1000 km apart, performed
controlled crosses, and grew the parents and progeny in common gardens. They found that
first-generation hybrids between plants from different populations outperformed their parents,
regardless of the geographic distance between sources. By the third generation, however, this
increase in fitness declined. The level of decline varied with distance between parent
populations, with crosses between plants from <1000 km apart yielding third-generation
Inbreeding and outbreeding in Castilleja levisecta, 2003
13
plants at least as vigorous as their original parents. Thus, crosses of up to 1000 km had a
short-term beneficial effect, and little long-term risk (at least through the third generation).
There have been too few studies of out-breeding depression to make generalizations about the
level of risk, however. Some studies have documented negative effects of outbreeding across
short distances (tens of meters to 100 m) (Price and Waser 1979, Waser and Price 1989, 1991,
1994) or between different habitats (Montalvo and Ellstrand 2001), while others have found
great variability in the effects of outbreeding, even in the same species (e.g., Waser et al.
2000). The threat of outbreeding depression is one argument against mixing seed sources
during plant restoration (Kaye 2001a). It is also one of the dangers of moving plants a great
distance to a restoration area where they could interbreed with a local population.
Outbreeding advantage may play a role in the formation of new species through hybridization
in Castilleja, which is a process that has been documented in the genus (Heckard 1962),
especially in the Intermountain West (Heckard and Chuang 1977). In addition, the presence of
hybrid swarms in mixed species populations of Castilleja may be facilitated by the lack of
inter-specific mating barriers (Anderson and Taylor 1983) and hybrid vigor of mixed species
crosses.
Recommendations for recovery and reintroduction
1.
If seed production is found to be low in any small, wild populations of Castilleja
levisecta, the intentional importation of pollen or transplanted individuals should be
considered to restore compatible mating types (or genetically unrelated individuals) to
the population. This potential recovery action should be evaluated along with other
explanations for low seed production, including insufficient pollination and resource
limitation (due to climatic effects, interspecific competition, or other factors).
2.
Seed collection for reintroduction projects should be managed to spread collections
among many individuals and emphasize non-neighbors, if possible. Neighboring plants
may be more genetically related than distant plants in the same population and more
likely to cross-pollinate with each other, thus increasing the potential for inbreeding
Inbreeding and outbreeding in Castilleja levisecta, 2003
14
depression. In addition, seeds from different individual plants should be stored
separately.
3.
Reintroduction projects should consider the genetic relationship of individuals in
founder populations and take steps to minimize the chances that members of the new
population are siblings or close relatives. Specifically, non-relatives should be planted
close to one another and relatives should be kept distant.
4.
Mixing plants from different source populations at reintroduction sites may improve the
fitness of their progeny, at least in the short term. At this time, this recommendation
should be considered only for adjacent populations in regions where the species is still
extant (e.g., Whidbey Island or San Juan Island) and when there is some suggestion
that the population is declining or is already very small, or in areas where the species
has been extirpated (such as the Willamette Valley) so that any risks of outbreeding
depression and local genetic contamination are minimized.
Inbreeding and outbreeding in Castilleja levisecta, 2003
15
LITERATURE CITED
Anderson, A.V.R. and R.J. Taylor. 1983. Patterns of morphological variation in a population of
mixed species of Castilleja (Scrophulariaceae). Systematic Botany 8:225-232.
Aspi, J. 2000. Inbreeding and outbreeding depression in male courtship song characters in
Drosophila montana. Heredity 84:273-282.
Caplow, F. 2002. Reintroduction plan for Castilleja levisecta (golden paintbrush).
Washington Natural Heritage Program, Washington Department of Natural Resources,
Olympia, Washington.
Fenster, C.B. and L.F. Galloway. 2000. Inbreeding and outbreeding depression in natural
populations of Chamaecrista fasciculata (Fabaceae). Conservation Biology
14:1406-1412.
Friar, E.A., D.L. Boose, T. Ladoux, E.H. Roalson, and R.H. Robichaux. 2001. Population
structure in the endangered Mauna Loa silversword, Argyroxiphium kauensis
(Asteraceae), and its bearing on reintroduction. Molecular Ecology 10:1153-1164.
Gharrett, A.J. 1999. Outbreeding depression between odd-and even-broodyear pink salmon.
Aquaculture 173:117-129.
Godt, M.J.W. and J.L. Hamrick. 2002. Allozyme diversity in the federally threatened golden
paintbrush, Castilleja levisecta (Scrophulariaceae). Unpublished report from the
University of Georgia, Athens, Georgia to the Washington Department of Natural
Resources, Olympia, Washington.
Guerrant, E.O., Jr. 2003. Recommendations for the reintroduction of Castilleja levisecta based
on patterns of genetic diversity as revealed by allozyme data. Unpublished report from the
Berry Botanic Garden, Portland, Oregon to the Washington Department of Natural
Resources, Olympia, Washington.
Inbreeding and outbreeding in Castilleja levisecta, 2003
16
Heckard, L.R. 1962. Root parasitism in Castilleja. Botanical Gazette 124:21-29
Heckard, L.R. 1968. Chromosome numbers and polyploidy in Castilleja (Scrophulariaceae).
Brittonia 20:212-226.
Heckard, L.R. and T.-I. Chuang. 1977. Chromosome numbers, polyploidy, and hybridization in
Castilleja (Scrophulariaceae) of the Great Basin and Rocky Mountains. Brittonia 29:159172.
Hufford, K.M. and S.J. Mazer. 2003. Plant ecotypes: genetic differentiation in the age of
ecological restoration. Trends in Ecology and Evolution 18:147-155.
Kaye, T.N. 2001a. Common ground and controversy in native plant restoration: the SOMS
debate, source distance, plant selections, and a restoration-oriented definition of native.
In, Native Plant Propagation and Restoration Strategies. R. Rose and D. Haase, editors.
Nursery Technology Cooperative and Western Forestry and Conservation Association,
Corvallis, Oregon.
Kaye, T.N. 2001b. Restoration research for golden paintbrush (Castilleja levisecta), a
threatened species. Institute for Applied Ecology, Corvallis, Oregon. Available online at
http://www.appliedeco.org/Reports/Cale_research.PDF.
Keller, M., J. Kollmann and P.J. Edwards. 2000. Genetic introgression from distant
provenance reduces fitness in local weed populations. Journal of Applied Ecology
37(4): 647-659.
Lynch, M. 1991. The genetic interpretation of inbreeding and outbreeding depression.
Evolution 45:622-629.
Meffe, G.K. 1996. Genetic and eco-logical guidelines for species re-introduction programs.
Journal of Great Lakes Research 21:3-9.
Inbreeding and outbreeding in Castilleja levisecta, 2003
17
Montalvo, A.M. and N.C. Ellstrand. 2001. Nonlocal transplantation and outbreeding
depression in the subshrub Lotus scoparius (Fabaceae). American Journal of Botany
88:258-269.
Price, M.V. and N.M. Waser. 1979. Pollen dispersal and optimal outcrossing in Delphinimum
nelsonii. Nature 277:294-297.
Quilichini, A., M. Debussche, and J.D. Thompson. 2001. Evidence for local outbreeding
depression inthe Mediterranean island endemic, Anchusa crispa Viv. (Boraginaceae).
Heredity 87:190-197.
Richards, C.M. 2000. Inbreeding depression and genetic rescue in a plant metapopulation.
The American Naturalist 155:383-394.
Rossiter, S.J., G. Jones, R.D. Ransome, and E.M. Barratt. 2001. Out-breeding increases
survival in wild greater horseshoe bats (Rhinolophus ferrumequinum). Proceedings of
the Royal Society of London Biological Society 22:1055-1061.
Sheridan, P.M. and D.N. Karowe. 2000. Inbreeding, outbreeding, and heterosis in the yellow
pitcher plant, Sarracenia flava (Sarraceniaceae), in Virginia. American Journal of
Botany 87:1628-1633.
Waser, N.M. 1993. Population structure, optimal outbreeding, and assortative mating in
angiosperms. Pages173-199 in N.W. Thornhill, ed., The Natural History of Inbreeding
and Outbreeding: Theoretical and Empirical Perspectives. University of Chicago Press,
Chicago.
Waser, N.M. and M. Price. 1989. Optimal outcrossing in Ipomopsis aggregata: seed set and
offspring fitness. Evolution 43:1097-1109.
——. 1991. Outcrossing distance effects in Delphinium nelsonii: pollen loads, pollen tubes,
and seed set. Ecology 72:171-179.——. 1994. Crossing distance effects in Delphinium
Inbreeding and outbreeding in Castilleja levisecta, 2003
18
nelsonii: outbreeding and inbreeding depression in progeny fitness. Evolution
48:842-852.
Waser, N.M., M. Price, and R.G. Shaw. 2000. Outbreeding depression varies among cohorts
of Ipomopsis aggregata planted in nature. Evolution 54:485-491.
Wentworth, J.B. 1998. Castilleja levisecta, a threatened South Puget Sound prairie species.
Pages 101-104 in P. Dunn and K. Ewing (eds.), Ecology and Conservation of the South
Puget Sound Prairie Landscape. The Nature Conservancy, Seattle, Washington.
U.S. Fish and Wildlife Service. 2000. Recovery plan for the golden paintbrush (Castilleja
levisecta). U.S. Fish and Wildlife Service, Portland, Oregon.
Inbreeding and outbreeding in Castilleja levisecta, 2003
19